The document discusses using surface electromyography signals from gesture movements for personal authentication. An experiment was conducted with 5 participants to measure electromyography signals from the forearm during 6 different gestures repeated 10 times each. The results showed similar signal waveforms for the same gesture performed by the same person, but dissimilar signals for the same gesture between people. Various machine learning methods were tested for gesture recognition and personal authentication using electromyography signals, achieving a false rejection rate of around 30% on average. Feature selection methods like ANOVA and random forests were able to reduce the number of important features needed for accurate authentication.